That's exactly what we're seeing (I'm also seeing some latency doing a save). I had someone from Tech Support tell me he had heard something out of the labs about a problem specifically with MS Office apps. Apparently, the Office apps send out a boatload of questionable RPC requests but ONTAP only responds to about 5 of them and Office waits until the others time out waiting for a response. He promised to send me anything written on it he could find but so far I haven't seen or heard anything else. We're still running Office 97 so I would have thought this issue would have come up and been solved a long time ago.
Michael
"Sayla, Mustafa" <MSAYLA@mesirowfina To: "'MCope@isisph.com'" MCope@isisph.com ncial.com> cc: Subject: RE: Performance issue 12/06/01 12:30 PM
We are also experiencing the same sort of problem. I have a F760 running 6.0.1R3. In our case the first time you open any office document it takes about 15-25 seconds but after that if you open the same document or any other document it opens in 2 seconds. We are also using NAV. If you have found out any solution please let us know. I also have a case open with Netapp for about a month now but no real solution from their side.
Mustafa
-----Original Message----- From: MCope@isisph.com [mailto:MCope@isisph.com] Sent: Tuesday, November 27, 2001 10:48 AM To: Haber, David J. Cc: toasters@mathworks.com Subject: Re: Performance issue
David,
Exactly what kind of issues are you seeing. I have the exact same setup except I have Gigabit instead of 100Mb. The issue primarily seems to be opening Office files from a share can take 30+ seconds. Other files types don't seem to be such an issue although I've personally seen a tiny text file take 20 seconds to open in Notepad. I've been the lone voice in the wilderness claiming it's not a Filer problem although it sounds like that might not be entirely true. We did see that running NAV with network drive checking enabled significantly made file access slower. We've done extensive testing and trouble shooting with the local NetApp folks. They discovered the NAV problem. The biggest reason I've not been looking at the Filer as a source of the problem is we've experienced the exact same behavior when accessing files from a share on an NT server. That in conjunction with some packet captures we've done seem to point to a problem with NT's TCP/IP stack. The stack starts out using very small packets and takes forever to ramp up to a decent packet size (Win2K is much faster - in tests a Win2K box on 100Mb ran laps around an NT server using Gigabit). We also use NIS as our primary means of authentication on the Filer and we still haven't ruled this out as an issue.
I'm certainly interested in learning more about your problem and everything you learn/hear. This has been made a "top priority" for me to resolve and it's been frustrating because there doesn't seem to be a "smoking gun" and in the meanwhile management and users are cursing the good name of NetApp.
Michael Cope Unix Systems Administrator Isis Pharmaceuticals mcope@isisph.com